Nomogram for predicting difficult transoral and submental thyroidectomy: a retrospective model development and validation study with large-scale population.

OBJECTIVE

No prior studies have described or stratified the difficulty of transoral and submental thyroidectomy (TOaST). We aimed to investigate preoperative factors as indicators of difficult TOaSTs and to develop a predictive model accordingly.

METHODS

This retrospective study included 255 eligible DTC patients who underwent total thyroidectomy and central neck dissection (CND) via transoral and submental endoscopic approach between February 2021 and April 2024. These patients were randomized into training and validation groups in a 7:3 ratio. Procedures were categorized into difficult and normal TOaST based on operation time, conversion to open and intraoperative injury. Univariate and multivariate logistic regression analyses were used to assess the association between surgical difficulty and factors regarding demographics, laboratory tests and ultrasound information. A nomogram was then developed and validated internally. Surgical and oncological profiles and follow-up data were also analyzed.

RESULTS

Five independent risk factors for difficult TOaST were identified in multivariate analysis: age (OR 0.84, p < 0.001), male sex (OR 4.75, p = 0.016), thyromental distance (TMD) < 7 cm (OR 7.59, p < 0.001), presence of diffuse changes on ultrasound (OR 14.5, p < 0.001), and elevated anti-thyroid peroxidase antibody (TPO-Ab) level (OR 5.22, p = 0.005). The nomogram performed well on both the training and the validation datasets, achieving an area under curve (AUC) of 0.908 and 0.888, respectively. Calibration curves for both datasets also fit well. There was no significant difference in complication rates between the difficult and normal TOaST groups.

CONCLUSION

The developed nomogram provides a reliable, straightforward prediction of difficult TOaST, thus supporting preoperative preparation and consultation, as well as optimizing training and promotion.